Background:  If tuberculosis therapy is to be shortened it is imperative that the sterilising activity of current and future tuberculosis drugs must be enhanced. Here we investigate whether Mycobacterium tuberculosis (MTB) generates tolerance to drugs over time, and whether alveolar macrophages provide MTB with a pharmacological sanctuary site.

Methods:  A modified microplate alamar blue assay was used to assess H37Rv drug susceptibility. Briefly, H37Rv (1x10⁷ colony-forming units) were incubated in the presence of rifampicin (0 to 166 ng/mL), isoniazid (0 to 5 µg/mL), and ethambutol (0 to 50 µg/mL) for 1, 2, and 3 weeks. Bacterial viability was then assessed by alamar blue turnover. Rifampicin- and isoniazid-tolerant H37Rv was selected from drug spiked Middlebrook 7H11 and grown in drug-containing or drug-free media (3 weeks) before their drug susceptibility was determined by microplate alamar blue assay.

The drug susceptibility of intracellular bacilli was determined by a novel method. Briefly, A-THP1 cells were infected with H37Rv (multiplicity of infection = 10) and grown in rifampicin (0 to 1 µg/mL), isoniazid (0 to 1 µg/mL), and ethambutol (0 to 10 µg/mL). Following incubation (1 week, 37^oC), A-THP1 cell viability was determined and used as an inverse marker of intracellular H37Rv viability.

Results:  Based on the microplate alamar blue assay there was a significant increase in rifampicin EC₅₀ from day 7 to 14 to 21 ([n = 4] mean 1.6 [0.7 to 2.0] to 7.3 [2.0 to 17.8] to 24.7 [5.6 to 45] ng/mL, respectively, p ≤0.05; paired student t-test) (figure A), and isoniazid EC₅₀ (47 [36 to 55] to 53 [35 to 71] to >5000 ng/mL, respectively, p ≤0.05). In contrast, ethambutol EC₅₀ did not change over time. Only selected rifampicin-tolerant, but not isoniazid-tolerant, H37Rv strains reverted back to wild type drug susceptibility following growth in drug-free media. The isoniazid-tolerant H37Rv showed no added tolerance to ethionamide. A higher concentration of rifampicin was required to kill intracellular (148±32 ng/mL) versus extracellular bacilli (1.3±0.02 ng/mL) (figure B). The intracellular drug activity of isoniazid (mean EC₅₀±SD, 37±2.2) and ethambutol (243±95) were similar to that of extracellular kill (57±2.5 ng/mL and 263±12 ng/mL, respectively).

Conclusions:  MTB becomes tolerant to rifampicin (phenotypic) and isoniazid (suggested genotypic) during 3 weeks of drug exposure. Furthermore, intracellular H37Rv rifampicin susceptibility is greatly compromised compared to that of extracellular bacilli. Both these factors may contribute to the slow clearance of MTB during therapy.